March 16

answer tool alkymi users will for the first time be able to benefit from cutting edge “answer tool” machine learning models in their business workflow the answer tool allows you to simply ask a question and a natural language machine learning model will search the document for the correct answer users will be able to see results regardless of any potential variations in document formats and contexts, making this a flexible tool to use when dealing with unknown document formats thanks to this flexibility, defining and revising this tool is fast and easy and maintains the traceability that other tools provide some examples of the question tool include unlocking key datapoints from equity research reports obtaining form data from a 1099 div we decided to build this feature because we wanted to give users flexibility in addressing different document formats and make our extraction methods more resilient to small changes for example, if the text surrounding a target value were to change slightly, a user would need to account for that when defining their text tool instead, we wanted to democratize ai by giving users an uncomplicated, no training required method by which to extract data